Why so quiet?

Short answer, I’ve been focusing a lot on my new job, with IntraPhone Solutions AB.

$DAYJOB

I’m realizing now that I had stagnated at my old job and was in desperate need for something new. I had been with Telia Cygate for 13 years! I met the love of my life there, and lost her to cancer while still working there.

At Telia/Cygate I leveled up a lot, learned so many new things, met so many talented people, it was truly a time of my life that I will never forget.

But the last few years I didn’t feel the spark anymore, I didn’t feel creative, or motivated. I thought I could get it back, I tried hard to get it back, but now I believe that I needed a change.

So I quit my job, and re-entered the job market in a brutal way, by being unemployed for a few months.

After several interviews, some job offers, I finally found IntraPhone.

I quickly fell in love with this job. They build everything with open source software, from firewalls, to hypervisors. They develop services for eldercare, and service a large number of Swedish municipalities. I love their mission, I love how close knit they are, like a little family of only ~25 employees, and I love that my dog is welcome at the office. After a year with IntraPhone I feel at home there in a way I haven’t felt in over a decade.

Most of the time I just want to feel like I’m making a positive difference in the world, and here I get that feeling. While also getting to work with open source, use all my talents when it comes to designing infrastructure, and get to work closely with a great group of down to earth people.

Website

Made a new design for my website but I forgot to fix the contacts section, it’s now fixed, with the correct public key for my e-mail address.

Refactored my homelab

My homelab was overly complicated with K8s, Ceph, on 6 nodes!

I sold some of the nodes and now I’m down to just 2 Asus Mini PCs with a combined 96G RAM, and Ryzen 5 and 7 CPUs, running RHEL10 with a bunch of rootless quadlets for services.

This, and my Synology NAS, is good enough for all my current homelab needs.

You can follow my homelab’s development at its Gitlab repo.

AI

Yes AI has been a huge part of my life too, it’s a very hot topic right now, let’s talk about it.

When it first came out I did use one to generate a couple of images for this blog, you can see them in old posts about Ansible, and Atomic Fedora. It was just an experiement from my side, I don’t approve of AI generated images when you could have hired an actual artist to create them, but I unfortunately see where this is headed and fear it will be the dominant method to generate any advertising graphic.

I always envied artists for their talent, personally I don’t have an ounce of graphical talent in my entire body. When tools like image generators come out, I see that as a salvation for people like me who are completely worthless at graphics. But in the larger context this will be detrimental to thousands of graphics artists out there.

And this was all I did with AI for several years after.

Like all new technologies I was skeptical, when I shouldn’t have been skeptical, I only started using a coding agent in the summer of 2025. Just like with containers, it took me 5 years before I started using containers, and now I can’t imagine life without them.

But unlike with containers, AI businesses are now gobbling up all RAM circuits in the world for their massive energy consuming datacenters. And this is only going to get worse, I can see the signs already. Everyone is using it, EVERYONE insert Gary Oldman. If they’re not using it to vibe code a new startup, or proofread their Dear John letters, they’re using it to create funny videos of cats playing instruments outside of people’s houses at night.

It’s clear to me that AI subscriptions are selling like hotcakes, so the RAM shortage issue is only going to get worse before it gets better.

I am convinced that during 2026 we will experience outages of critical services because they were unable to scale or repair their infrastructure, due to the RAM shortage.

That said though, from a personal perspective, AI has absolutely made my life better in the short run.

It’s a very powerful tool that can generate a lot of code for you in a short time, but like all tools you must know how to wield it, to get the best results.

The way I use AI is almost like I’m the solutions architect, and AI is the developer. This means that I would never attempt to use AI for something I couldn’t do myself.

That is the most important guideline I have discovered, NEVER HAVE AI DO SOMETHING YOU COULDN’T DO YOURSELF.

So the term vibe coding to me actually originates from people who have no previous experience, and jump head first into creating an entire project, in a language they don’t understand.

I would never create a project in Rust, or Golang, because I’m not proficient in those languages. I wouldn’t be able to fully understand what the AI was generating for me, and that is a recipe for disaster.

So I don’t really view what I do as vibe coding, it’s just a tool to generate code for me. We have had tools that generated code before, it has just gotten more advanced.

And this brings me to the next general guideline I have made for myself regarding AI, A PERSON IS STILL RESPONSIBLE FOR ALL CODE SUBMITTED.

It doesn’t matter if you had AI generate your entire Merge Request, or your entire Bug report, YOU are still the responsible person that submitted it. If you submit garbage to a project or a maintainer, you will be banned from submitting anything else, and that is your responsibility, not the AI’s.

AI generated music

This is my big pet peeve, sure coding agents are making my life easier, but AI generated music does absolutely NOTHING for me. And as we all know, it’s all about what it does for me.

I enjoy an eclectic selection of music, but my main genre is the 90s hip-hop of my youth. And AI generated music has made me realize that my favorite music is 25% image, and only 75% skill.

That is one thing AI can never replicate, the emotion that can only come from expressing your own unique life and experience into your music.

So I don’t think it’ll ever impress me, but I am seeing people around me get into AI generated music. I can tell where the winds are blowing, and it’s not sounding good to me.

It seems to me that in a near future real music will be something only hipsters like me and my friends listen to, and a large majority of people will be satisified with some percentage of AI generated music.

A lot of cookie-cutter pop-music, break room music, work site music, will be replaced by AI generation. People who never really cared who the artist was, and focused more on the music of the artist.

That’s the big irony of hip-hop, MF DOOM tried to make us stop focusing on the artist, but 50% of the reason we love MF DOOM is because he was such a unique person, with amazingly unique rhymes and productions.

AI can never be a unique person, because that comes from struggle and experience.

So yeah, those are my 2 cents on AI at this moment…